Henry Melson Stommel<Br> 27 September, 1920Œ17 January
Total Page：16
File Type：pdf, Size：1020Kb
Journal of Marine Research, 50, i-viii, 1992 Henry Melson Stommel 27 September, 1920-17 January, 1992 Henry Stommel's heart stopped beating shortly after midnight on Friday, January 17, 1992, four days after he had undergone surgery for liver cancer at Deaconness Hospital in Brookline, Mass. His death brought to an end the career of a man who, for 45 years, was the most significant scientific contributor to the development of oceanography and who brought a rare degree of harmony and collegiality to the field. When Hank arrived at the Woods Hole Oceanographic Institution in 1944 there was little reason to suppose that anything momentous had taken place. As an undergraduate at Yale he had been advised by a counselor that, since he evidently had no talent for science, he should take up law. In 1944 he was a second-year graduate student in Astronomy at Yale and was a conscientious objector to war. The job at Woods Hole was a way of serving his country without going to the battlefield. The Yale Astronomy Department from which he had come was strongly focused on celestial mechanics, and Hank had developed an interest in the marine environ- ment through his study of celestial navigation, one of the courses that he taught. He had read a lot about the ocean and decided to prepare a synthesis that he dedicated to the students in the navy program in which he had been teaching. It was a trait that was to stay with him thoughout his life; he would be heard. Over a three-week period he wrote a 208-page book, Science of the Seven Seas, which was published by Cornell Maritime Press in 1945. In later years he was somewhat embarrassed about that qualititative survey, but it sold more copies than all of his other books combined. Toward the end of the war he moved in with a group of WHOI bachelors who occupied the former Episcopal rectory in Woods Hole. His clowning around during those bachelor years was legend and so was his intense study of all aspects of oceanography. He liked people and he loved life and the wonders of nature. In the years that followed he took part in many oceanographic cruises where he could enjoy the social and professional camaraderie of his colleagues in a setting devoid of the distractions of modern life. Life at the rectory served as an introduction to life at sea. Hank's professional development during those early years resulted in an intriguing mix of publications. Some of them, summarizing what he had learned, amounted to a kind of report card. They were part of what he would later refer to as his "commentarial ambition." Others, reporting measurements and observations, docu- 11 Journal of Marine Research [50, 1 mented his growing familiarity and involvement with the world of nature. But it was the third category that revealed his originality and true genius. One of his first articles, in 1947, dealt with the computation of the dynamic height anomaly, a topic that was anything but arcane. He proposed using the observed T-S correlation in certain regions to estimate the missing salinity when only temperature had been measured, and thereby to determine an approximation to the dynamic height anomaly. What makes that study conspicuous is not the careful evaluation of the feasibility of the method but the issue that he addressed. Even in his first year of published research papers he was already thinking about how to make up for the inadequacy of oceanic observations that has always plagued physical oceanography. Two years later he considered a steady convection cell with upward flow in the middle and downward flow at the sides and asked whether particles heavier than water could remain in suspension. He proved that as long as the terminal velocity of the particles is smaller than the maximum fluid velocity there would be a closed region containing suspended particles. Since 1949 people in many different fields have come to the same conclusion. Though he felt that the result would be applicable to cloud physics and to a suspension of particles or organisms in the sea, it was the study of the physical process that intrigued him. Of course, it was the westward intensification paper that put the name, Stommel, permanently into the dictionary of ocean circulation. In 1955, when he was invited by Jule Charney to visit the Institute for Advanced Study in Princeton, he and I worked on a problem together, and I asked him whether the westward propagation of Rossby waves suggested the idea to him. He said that there was no connection and told me the following story: Early in 1947 Ray Montgomery mentioned to him a question that had been posed by Columbus Iselin, who was then Director of WHOI. Why was the Gulf Stream a narrow current pressed against the western side of the Atlantic when neither the thermal driving nor the wind had any such asymmetry? Hank set up a theoretical model to treat the problem and found that he had to solve a boundary value problem involving an elliptic partial differential equation. He had derived another such equation for a tidal problem and had learned that R. V. Southwell had just published a book on relaxation methods for the numerical solution of elliptic equations. So he decided to learn the relaxation method by applying it to the circulation model. This was before electronic computers were developed; relaxation by hand involved many months of tedious computations using a mechanical calculator. Since the wind stress in his model had east-west symmetry, his initial guess for the circulation had the same symmetry. But in reducing the residuals he found that the circulation immediately started to shift toward one with a stronger current on the western side. After some experimentation he discovered that the westward shift was caused by the beta term (the variation of the Coriolis parameter with latitude). So instead of continuing with the relaxation method, he formulated a simpler model, one that contained the beta 1992] Henry Melson Stommel 1\1 term but could be solved analytically. The publication of that analysis in the 1948 Transactions of the American Geophysical Union marked the beginning of modeling of large-scale ocean circulation. Hank said that that experience taught him the value of developing models that isolated the essential physics in the simplest mathematical context. As the years wore on, he became so proficient at that and developed such a powerful intuition that in starting a discussion he would leap past all of what he considered unessential preliminaries and start with the analysis of his "simple" model. Not infrequently the listener would just hang on, hoping that as Hank neared the end, the original question might become apparent. Carl Rossby, who was then at the University of Chicago, met Hank in Woods Hole just after the end of the war and invited him to visit Chicago in spring, 1946. During those postwar years and even up to 1960 when he was appointed Professor at Harvard, Hank bemoaned his lack of a Ph.D. His visit to Chicago was an exploration of possible graduate study there. After delivering his westward intensification paper at the AGU meeting in Woods Hole in September, 1947, he left for England on a six-month leave of absence without pay. From his meager salary (he received $1300/year initially) and the royalities from Science of the Seven Seas he had saved $1500 to finance the trip. His initial purpose was to visit Southwell and become really familiar with the relaxation method, but the temperaments of the two men were so different, that Hank soon abandoned that idea and spent his time with the group working on meteorology and oceanography at Imperial College in London. His most enjoyable visit was with another pacifist, L. F. Richardson, who had retired to Scotland. The two hit it off immediately, and even though the visit was short, it resulted in a joint note on turbulent diffusion, reporting their observations of the relative motions of pairs of parsnips thrown into Loch Long from a pier. On his return to WHOI Hank extended his areas of research activity by contribut- ing to a lengthy treatise on the ecology of the Atlantic and another on atmospheric observations over the Caribbean Sea. He then spent several months at Scripps in 1949. It may have been during that trip or sometime earlier that he probed the possibility of graduate study at Scripps. He was not encouraged. He told me not long after we met that he thought that H. U. Sverdrup, the Director of Scripps, disapproved of his watered-down treatment of oceanography in Science of the Seven Seas. In any event, after that trip he never seriously considered graduate study again even though he continued to feel uncertified. In 1950 as Stommel's westward intensification model was gaining recognition, Ray Montgomery moved to Brown, where he occupied an office next to the Applied Mathematics Department. He invited various physical oceanographers, Stommel and Walter Munk among them, to visit and within months Stommel's westward intensification model was reinterpreted by George Carrier in terms of a viscous iv Journal of Marine Research [50,1 western boundary layer and an inviscid interior. That separation of the flow into two regions that could be joined caught on quickly and in practically no time Stommel's model was regarded as providing the viscous closure in an enclosed basin of the Sverdrup planetary vorticity balance. That is a logical fluid dynamical interpretation but the Sverdrup relation and Stommel's model were derived completely indepen- dently.